hrberg has quit [Quit: No Ping reply in 180 seconds.]
ec has quit [Remote host closed the connection]
hrberg has joined #commonlisp
ec has joined #commonlisp
bilegeek has joined #commonlisp
waleee has quit [Ping timeout: 268 seconds]
waleee has joined #commonlisp
jeosol has joined #commonlisp
gko`` has quit [Quit: ERC 5.4 (IRC client for GNU Emacs 28.2)]
<jeosol>
good morning all!
gko has joined #commonlisp
<jeosol>
any update on process to update quicklisp dists, what is the frequency? I was trying to update to sbcl 2.3.4, with causes an error with log4cl. A search online reveals this was also present in 2.3.3. Did anyone remember encountering this issue?
<jeosol>
For context, I am trying to move from 2.3.2 to 2.3.4.
* Lycurgus
vaguely remembers something about this
<Lycurgus>
the solution is don't update ql
<waleee>
newer versions (ie not the one in quicklisp) works with new sbcl
<waleee>
*the
<waleee>
or well, it worked with the a git version of sbcl master from about a week ago at least
<waleee>
s/the//1
* waleee
updates his sbcl
Lycurgus has quit [Quit: Exeunt: personae.ai-integration.biz]
tyson2 has quit [Remote host closed the connection]
<jeosol>
lycurgus: haha, that's funy, thanks for the laugh. I haven't updated ql since feb, but it seems the with recent sbcl, something is broken.
<jeosol>
Usually, I try to stay with a stable ql and sbcl combo and stay there, but at the same time, I don't want my codebase to lag too much so the upgrades won't be too much work. Usually, I stay 2 versions behind latest sbcl.
<jeosol>
walee: you mean with the lastest ql? I think the latest sbcl is April 30
<waleee>
I git cloned log4cl to ~/quicklisp/local-projects which afai tell seems to work fine with the rest of the current quicklisp dist
<waleee>
fun. Now one of the libraries (not log4cl) that nyxt uses didn't build
<jeosol>
walee: thanks for clarifying that. Make sense. I was building with docker and wanted it to pull using a set ql. I can update my local version using a locally cloned log4cl. Usually, I keep local and docker images sync.
<jeosol>
I did a search online and I think someone found out that ultralisp does pull the latest log4cl.
dcb has quit [Quit: MSN Messenger v2.16.1]
drainpipe has joined #commonlisp
azimut has joined #commonlisp
dtman34_ has quit [Ping timeout: 248 seconds]
waleee has quit [Ping timeout: 268 seconds]
szkl has joined #commonlisp
<skin>
This touches on a question I've had for a while
<skin>
For dependencies I use qlot
<skin>
But everyone seems to be perfectly content with QL which is rolling release
<skin>
I've never seen that before in a language community it's puzzling
<skin>
jeosol: if you're looking for a way to nail down dependencies or use a particular get version I find qot to be very nice
<skin>
git version not get version
<skin>
qlot not qot
<beach>
skin: What was your question?
<skin>
The question was why everyone's okay with the rolling release of QL.
<skin>
There's no way to pin versions down. No version numbers even.
<beach>
I am guessing, but there are probably very few people wanting to do the job that Xach is doing, so they are OK with whatever he does.
<skin>
I hear that
<beach>
He got a standing ovation at ECLM (in Amsterdam?) when he announced Quicklisp.
<skin>
What year was this
<skin>
Super interesting history
<skin>
Very intriguing. I use q lot right now which builds on quick lisp
<beach>
I forget. More than 10 years ago.
<skin>
Most of the stuff I get from it is built on it.
<skin>
Ok
<skin>
What is eclm
<beach>
European Common Lisp Meeting, no longer exists. It was organized by Edi Weitz and Arthur Lemmens.
<beach>
But Edi got busy with a steady job, so they stopped organizing it.
<beach>
It was not an academic conference, and speakers were invited.
drainpipe has quit [Quit: WeeChat 3.8]
<beach>
I guess 15 years ago, because I think it was the same year as the first ELS.
<edwlan[m]>
skin: pinning versions is only really necessary because packages in other systems break regularly
<edwlan[m]>
I find packages in CL stable enough that I can handle the occasional breakage in a rolling-release model
<edwlan[m]>
And, having the distribution mechanism be like this makes people less likely to release breaking changes imo
<edwlan[m]>
The other factor here is that I typically end up loading all my projects into a single running lisp image over time
<skin>
Sort of like Java then
<skin>
Their package manager also encourages them to release backwards compatible changes
<edwlan[m]>
Maven does version pinning, though
<skin>
Pretty cool
<skin>
True but not by default
<edwlan[m]>
And, in practice, any old java package is a nightmare to deal with updates
<skin>
Haha
<edwlan[m]>
Yes by default, it's just a different algorithm from modern systems like npm
<edwlan[m]>
node can load multiple versions of one package, mvn can't by default
<mariari>
CLPM lets you deal with different versions of packages at once between projects right?
<edwlan[m]>
There's no way to load multiple versions of a package into a single lisp image
<edwlan[m]>
s/package/system/
<beach>
Right, because of the lack of first-class global environments.
<skin>
Just another thing that forces people to play nice and write good code
<skin>
Honestly NPM is a nightmare
<skin>
You can get any version you want so you do and that people end up downloading 40 different versions of the same thing that's deep down the dependency stack
<skin>
And then they hit broken diamond dependency problems
<edwlan[m]>
I think this is a interesting case where the things various systems make easy have interesting outcomes
<skin>
Exactly
<edwlan[m]>
npm really solves the "multiple versions of a system/package" well
<edwlan[m]>
But it turns out that solving that problem well makes the ecosystem a bit annoying
<mariari>
beach: I wonder after your first class global environment changes, would it satisfy Joe Armstrong's criteria, I assume you'd want to have more crash isolation between different loaded programs in the SICL image
<beach>
mariari: I think that's partly a safety issue. A good Common Lisp system should not "crash". A thread might signal an error, but that should not influence the rest of the system.
<mariari>
beach: I agree but I have seen for example me going to ldb killing the image, or nyxt killing the image due to webkit below it mangling data, etc
<mariari>
it would be nice to get erlang style isolation, so I can just run most of my lisp programs in one image, and have them offer advanced services to other parts of the system
<beach>
mariari: I don't see how to do that. If some code corrupts the image, then all bets are off, and it would take a complete violation of Common Lisp semantics to avoid that, I would think.
<beach>
... as in objects being copied, etc.
<mariari>
I've gotten into ldb in SBCL by dumping docs with mgl-pax enough times
<beach>
mariari: I don't know what mgl-pax is, but I have two remarks: 1. SBCL puts performance first, and safety not so much. 2. If you run foreign code, all bets are off.
<beach>
I see running foreign code in a Common Lisp system the same way as I see running kernel modules in an operating system.
<mariari>
it could be very well a non-issue for most applications, if the LDB situations I've gotten into were due to that fact, then it wouldn't prevent a more wide use of a single image on the system with services
dtman34 has joined #commonlisp
<beach>
It could be due to bugs as well, of course.
<beach>
But I mean, if your operating system crashes, you don't call for crash isolation; you want the code to be fixed.
<beach>
But the prevailing attitude seems to be that it's OK for a programming system to crash but not for the operating system. I see no great difference between the two.
<edwlan[m]>
I think operating systems typically have some mechanisms for fault isolation, though
<mariari>
I don't think the language system ought to crash so easily, much like an operating system. However if one writes code to deliberately make it crash then so be it (in smalltalk you can invoke this with true become: false)
<edwlan[m]>
e.g. if an audio driver or something fails, I don't necessarily want my entire system to go down with it
<beach>
edwlan[m]: If you run a kernel module that corrupts the code for (say) the file manager, I think you are out of luck.
<edwlan[m]>
Linux at least has some degree of isolation between modules
<edwlan[m]>
And nice failure modes for particular classes of crashes
<mariari>
edwlan[m]: I've had bluetooth module on linux actually corrupt being able to update my system and do any writes to the rest of the system
<beach>
mariari: Running foreign code is in that category for me.
<edwlan[m]>
I've definitely had that sort of failure too
ns121 has joined #commonlisp
dirtcastle has quit [Ping timeout: 250 seconds]
<mariari>
had a scary situations where the kernel image didn't get rebuilt as it was painstaking to even get root to try to fix as the bluetooth module somehow made that operation have massive faults
ns12 has quit [Read error: Connection reset by peer]
ns121 is now known as ns12
<mariari>
beach: by foreign code you mean any code you did not write or any code that calls into C?
<beach>
Any code that is not Common Lisp.
<beach>
Usually C.
semarie has quit [Ping timeout: 248 seconds]
semarie has joined #commonlisp
<beach>
C and Common Lisp are both inherently unsafe languages, but: 1. We have a SAFETY OPTIMIZE quality and 2. The traditions are different. In C, the tradition is to use the unspecified parts of the standard for maximum performance, no matter the safety. In Common Lisp, the tradition is to avoid crashes in "safe code".
dirtcastle has joined #commonlisp
<beach>
The WSCL project aims to specify what most Common Lisp implementations already do, namely they avoid crashes in safe code (SAFETY 3).
<jmercouris>
well, we have to make compromises to integrate with the rest of the world...
<beach>
I don't.
<jmercouris>
I'm glad for you :-D, I wish I could say the same, but it would be in conflict with my objectives!
<jmercouris>
With regards to Quicklisp, yes... we have abandoned it some time ago in favor of Git submodules pinned to exact commits and versions.
<beach>
I think many of those compromises are due to the fact that we lack essential software written in Common Lisp. But it looks to me like most people then compromise by using foreign code, rather than helping with the development of the lacking software. Developing the lacking software would help others to use safe software, but instead, everybody struggles with foreign code.
jonatack1 has quit [Ping timeout: 268 seconds]
jonatack1 has joined #commonlisp
<beach>
I guess what we are observing is a variation on the prisoner's dilemma.
<beach>
It would be globally better for everyone to collaborate on developing the lacking software, but for each individual, it is easier to go the FFI way.
<beach>
jmercouris: Surely, I can't imagine your objective is to write unsafe code, right?
<jmercouris>
beach: no, it's definitely not to write unsafe code
<jmercouris>
but I also don't have millions of dollars in order to rewrite a web renderer, but soon I will ;-)
<beach>
I have had the impulse to as the EU for a multi-million-euro grant to develop Common Lisp software, but I have this feeling it would be in vain.
<beach>
*to ask the EU
<beach>
I wouldn't take any of the money myself, but I think I know a few people I would hire.
jmdaemon has joined #commonlisp
szkl has quit [Quit: Connection closed for inactivity]
<hayley>
My favourite coauthor and I had pondered getting a (much smaller) grant for developing our capability-safe language, but I'm not too sure who to ask.
<beach>
I am sure the Australian government has programs for things like that, no?
<hayley>
I don't get the impression the Australian government is very good around security, though maybe things have changed. Recently several large companies were hacked; I at least know that the way Optus (an internet service provider) was hacked was especially stupid. CSIRO (the national research agency) also kicked out the developers of seL4 in 2021, though apparently the developers had a bad relationship with CSIRO before.
<hayley>
Seems there is such a program, still.
chrcav has quit [Ping timeout: 268 seconds]
Cymew has joined #commonlisp
<beach>
Can't hurt to try.
<beach>
.. other than the time spent, of course.
rgherdt has joined #commonlisp
chrcav has joined #commonlisp
kenran has joined #commonlisp
<jackdani1l>
uh, so much backlog
pve has joined #commonlisp
<jackdani1l>
"Foreign function not found: X86-LINUX64::|memset|" hmm, I seem t orecall this issue on ccl
<splittist>
Is there a list of 'unawesome' 'scientific' 'journals' so I can not waste my time reading articles published in them?
<splittist>
"a text document is organised in a hierarchical structure in a top-down passion". Is it too much to expect NLP papers to use, you know, NL?
<jackdani1l>
I thought that you want something unawesome :)
<splittist>
"The ... is an advanced that feature heavily rely text mining techniques to extracts the headings and subheadings to be used for the constructing the ..." (sic). This is incorrect on every level. And from an English 'University'.
<splittist>
IJSCI International Journal of Computer Science Issues. I can certainly see lots of issues with computer science here...
<jackdani1l>
let's be honest, if /you/ are working on nlp then /you/ probably have some problem you want to automate
<jackdani1l>
i.e grammar
<beach>
splittist: A while ago, many journals eliminated their copy editors, and also a while ago, some of them publish pretty much anything that anyone is willing to write. It is very sad.
<beach>
*some of them started publishing
<jackdani1l>
aren't journals organizations that take a hefty price for access so they can offset their editorial costs?
<beach>
No, it's so that they can make more profit.
<beach>
Books about science and technology are very bad as well, not only at the language level, but with respect to the contents as well, as I have pointed out several times in the past. Like the OS books, for instance.
<splittist>
Obviously a far greater share of society's resources than is justified are being expended on the education-industrial complex.
<jackdani1l>
why oh why asdf/ccl can't follow symbolic links in my common-lisp directory? :( it works fine with asdf/{ecl,sbcl}
jackdani1l is now known as john_titor
<david`>
I'm surprised it doesn't. I'm trying to remember what I do in such situations (it's me - hhdave BTW, just haven't sorted my nick yet)
Algernon69 has joined #commonlisp
<john_titor>
I think that it is something about asdf building its registry and not really following symbolic links in its directory walker
<david`>
For a long time I used to use a (I think it was called) link-farm directory for ASDF in CCL. I just had a directory with ASDF systems in and I would link the */*.asd files into that directory. I realized that was unnecessary with ~/common-lisp though. I guess that's different from what you're trying to do though.
<john_titor>
perhaps I'll slap there directly the asd file then, thanks for the hint
<john_titor>
on the other hand that will be rather mundane, because mcclim has quite a few asd files
<john_titor>
so maybe I should just move the directory instead of linking it
<david`>
I had a simple command to just link to all the asds. I feel CCL should just be following the links though. What exactly is your setup?
<john_titor>
nothing fancy, vanilla ccl and quicklisp with "repo" directory linked in ~/common-lisp
<david`>
so you have ~/common-lisp/some-repo --> /some/path/some-repo as a symlink?
<john_titor>
yes
ahc has quit [Quit: Client closed]
Brucio-61 has quit [Remote host closed the connection]
Brucio-61 has joined #commonlisp
Algernon69 has quit [Ping timeout: 246 seconds]
<david`>
so it seems the behaviour of #'subdirectories in the :uiop package varies between CCL and SBCL. The latter returns a directory for a symlink but CCL doesn't.
<john_titor>
well, I've simply copied the directory, too many things going on to tackle ccl issues at the same time. that said - McCLIM now works fine with CCL when multiple threads run different frames
<david`>
That sounds good :)
<john_titor>
the issue was caused by the fact that ccl defaults a stream sharing mode to :private, so it errored when streams for TTF fonts were used between different frames
<david`>
Oh I see. Yes it does. I encountered that while trying to control lego robots from CCL.
<john_titor>
right, so now I'm opening the stream myself and passing it to zpb-ttf (instead of passing the filename)
<john_titor>
and when opened by ccl, it uses the shared mode
Algernon69 has joined #commonlisp
<david`>
OK. I'll leave that symlink issue for now, but at least I (or someone else) will know where to start looking to fix it.
Algernon69 has quit [Ping timeout: 264 seconds]
Guest86 has joined #commonlisp
Guest86 has quit [Client Quit]
Gleefre has quit [Remote host closed the connection]
Gleefre has joined #commonlisp
kevingal has joined #commonlisp
Lycurgus has joined #commonlisp
Lycurgus has quit [Client Quit]
kenran has quit [Remote host closed the connection]
karlosz has quit [Quit: karlosz]
Cymew has quit [Ping timeout: 240 seconds]
Krystof has quit [Ping timeout: 260 seconds]
kevingal has quit [Ping timeout: 252 seconds]
kevingal has joined #commonlisp
kevingal has quit [Ping timeout: 240 seconds]
kevingal has joined #commonlisp
random-nick has joined #commonlisp
Brucio-61 has quit [Ping timeout: 260 seconds]
Brucio-61 has joined #commonlisp
ahc has joined #commonlisp
tyson2 has joined #commonlisp
Algernon69 has joined #commonlisp
Algernon91 has joined #commonlisp
Algernon69 has quit [Read error: No route to host]
Algernon91 has quit [Ping timeout: 248 seconds]
rgherdt has quit [Ping timeout: 250 seconds]
rgherdt has joined #commonlisp
<splittist>
Just in case anyone wondered, the paper "SED: An algorithm for Automatic Identification of Section and Subsection Headings in Text Documents", IJCSI vol 17 issue 6, November 2020, doi 10.5281 zenodo.4431057 with four authors from the School of Computing and Mathematics, Coventry University is beyond trash. I feel stupider and dirtier for having read it. I am only sparing the authors' blushes by not mentioning them by name
<splittist>
because they might have moved on from publishing this stuff to more honest activities like drug dealing or pornography.
<Equill>
Uh, noted.
<yitzi>
I especially love the "Algorithmic Design" section. Brilliant logic! ;)
ahc has quit [Quit: Client closed]
Krystof has joined #commonlisp
<flip214>
splittist: what are your goals w.r.t. text analysis?
<splittist>
flip214: I have so many (: One, which is probably a perfect match for regexps (not much of a pun intended), is to detect "manual" section/clause headings (i.e. where someone has typed-out "2.3 Foo Bar" rather than relying on the automatic '2.3' numbering).
<prokhor>
splittist: thx for my first LOL moment today! :)
Cymew has joined #commonlisp
<splittist>
prokhor: glad to have been of use.
<splittist>
I was going to leave it there, on the grounds that the paper was no doubt th
<splittist>
e work of four undergrads and published on a lark/by accident/because of some weird idea of their teacher. But it seems to be an extract from a successful (?!) PhD thesis, the authors being the new doctor and his three supervisors.
<john_titor>
maybe you are being part of some experiment where they give you absurd data and see how you'll react? :p
<bjorkintosh>
hadn't seen that posted before now. never even seen it before
<flip214>
splittist: ah, okay
Krystof has quit [Ping timeout: 240 seconds]
Lycurgus has quit [Quit: Exeunt: personae.ai-integration.biz]
bjorkintosh has quit [Remote host closed the connection]
bjorkintosh has joined #commonlisp
bjorkintosh has joined #commonlisp
bjorkintosh has quit [Changing host]
<splittist>
flip214: actually, the more pressing is detecting when a document makes a structural shift, such that the previous heading is no longer applicable. To use a bad example, when the ‘Summary’ turns into the endnotes, you don’t want to refer to the third endnote as the sixth paragraph of the ‘Summary’ section.
<splittist>
Well, I don’t want to.
<splittist>
ObLisp, this is in the context of claarity (GitHub.com/splittist/claarity)
attila_lendvai has joined #commonlisp
cage has joined #commonlisp
jonatack1 has quit [Quit: WeeChat 3.8]
jonatack has joined #commonlisp
hineios has quit [Ping timeout: 240 seconds]
hineios has joined #commonlisp
Brucio-61 has quit [Ping timeout: 260 seconds]
Brucio-61 has joined #commonlisp
tyson2 has quit [Remote host closed the connection]
karlosz has joined #commonlisp
attila_lendvai has quit [Read error: Connection reset by peer]
<NotThatRPG>
I feel like I should know an answer for this, but I don't -- is there a way to mock functions for testing? The particular use case: alexandria is incorrectly computing its second return value that is derived from successive calls to `subtypep`. The values returned from `subtypep` vary across lisp implementations. So to make portable regression tests, I would like to make `alexandria:type=` invoke a mock version of
<NotThatRPG>
`subtypep` in the bounds of a test.
<NotThatRPG>
I don't believe that there's a portable advise function, which would be the obvious way to do this.
<NotThatRPG>
I suppose I could set the symbol-function value of subtypep, but I suspect implementations with package locks would refuse to do that.
karlosz has quit [Quit: karlosz]
<john_titor>
avoid cases where implementations are unsure maybe
<NotThatRPG>
Yeah, this stinks. RT is really a very rudimentary testing system.
<NotThatRPG>
john_titor: I suppose, but that requires me to find 4 testable examples and check them on all the implementations. And it seems like the implementations vary pretty widely
<john_titor>
no no, I mean
<NotThatRPG>
I would need to find four cases that the implementations agree on that will return (t t) (t nil) (nil nil) and (nil t)
<john_titor>
(is (or second-value-nil typy=)
<john_titor>
)
Krystof has joined #commonlisp
<john_titor>
you are not optimizing these tests for performance so you may very well inline the test for applicability
<NotThatRPG>
No, it's worse than that. I need to find cases which give all combinations of 16 values for (subtypep a b) and (subtypep b a). That's really not feasible or at least I refuse to do that myself.
<NotThatRPG>
I'm trying to verify that alexandria computes the second value correctly (which it does not in some cases).
Brucio-61 has joined #commonlisp
* john_titor
shrugs
<NotThatRPG>
So looking for a test that reliably fails today and will succeed after I fix alexandria:type=
<NotThatRPG>
RT is simply too crude a testing framework for this task. I could replicate the logic from alexandria:test= into the test, but that would no longer be testing the code, and any future changes to type= would go untested.
<john_titor>
you could test the second value only if that's the bug
shka has joined #commonlisp
<john_titor>
and throw at it randomized types or whatever rt defines in the universe
<john_titor>
that's hardly replicating logic if you assert that two T gives T in test=, any NIL gives NIL
<beach>
It looks to me lie the last example contains a mutation of a literal list. No?
<john_titor>
beach: yes
<beach>
Thanks.
<john_titor>
sure
<NotThatRPG>
john_titor: Yes, but my problem is that the implementations vary widely on what they yield for the second value of subtypep, making it hard to find a set of parameters for type= that will provide a test on all implementations.
* NotThatRPG
is crabby about having found a behavior that's trivial to achieve in python, but prohibitively difficult in CL
<john_titor>
if you define the test in term of the subtypep result (not a hardcoded guess) then you are all set
<john_titor>
and you test only a small subset of test= functionality
czy has joined #commonlisp
<NotThatRPG>
john_titor: Yes, that is what I am trying to achieve, but I don't see how to inject subtypep results into the body of type=
<john_titor>
well, you don't compare with pypi and cython
<john_titor>
you misunderstand - compute the second value baswd on manual subtypep at runtime
<john_titor>
and compare with 2nd value of test=
<john_titor>
I need to go, laters \o
<NotThatRPG>
john_titor: If I understand correctly, you are suggesting doing this randomly, right?\
<john_titor>
either that or against a product of *types* x *types*, rt has some var like this
markb1 has quit [Ping timeout: 260 seconds]
david` has quit [Remote host closed the connection]
* NotThatRPG
remembered wrong: RT does have an expected failure facility...
<NotThatRPG>
(to be fair, searching for "expected" in rt-doc.txt reveals no discussion of said facility)
tyson2 has joined #commonlisp
Algernon91 has joined #commonlisp
Algernon91 has quit [Client Quit]
<NotThatRPG>
One could make typep= look up its subtypep function (using alexandria::subtypep, e.g.) instead of just using cl:subtypep, but that doesn't seem very nice...
markb1 has joined #commonlisp
<edwlan[m]>
NotThatRPG: would it be possible to assert equality between the second return value of a call to subtypep and the second return value of the function in Alexandria?
cage has quit [Remote host closed the connection]
cage has joined #commonlisp
<edwlan[m]>
I see john_titor already said this
<edwlan[m]>
Something I’ve occasionally thought would be useful would be to use an alternate CL package based on features
<edwlan[m]>
#+dev :cl-advised #-dev :cl
<edwlan[m]>
Where the new package just wraps as much as possible from the old package and adds hooks for things like this
xaltsc has joined #commonlisp
slightlybrkn has quit [Remote host closed the connection]
ym has joined #commonlisp
taichi has joined #commonlisp
mariari has quit [Ping timeout: 265 seconds]
tyson2 has quit [Remote host closed the connection]
rgherdt has quit [Remote host closed the connection]
rgherdt has joined #commonlisp
<NotThatRPG>
edwlan[m]: That seems like a good idea (sorry for slow answer, I was AFK). The problem with using the alternate CL package based on features is that RT testing is "in image" --- the tests aren't run in a different lisp image. I suppose one could force ASDF:TEST-SYSTEM to recompile to get the new features, but then one would be left with the wrong version of the system at the end of the testing process...
<NotThatRPG>
It seems like one would like a method that forces the lisp implementation to find a function (in this case subtypep) by lookup through the symbol at invocation time. That would allow the tests to add a shadowing subtypep symbol into the alexandria package temporarily. But I don't know how to do this portably
<NotThatRPG>
Is there a way to determine if a function has been inlined?
kevingal has quit [Ping timeout: 268 seconds]
vassenn has joined #commonlisp
<jcowan>
splittist: I think "passion" is an error for "fashion", which suggests that the paper was dictated.
<edwlan[m]>
NotThatRPG: my thought about the alternate package would be to basically use it all the time when developing
<edwlan[m]>
But this is a bit harder for Alexandria, I guess
<NotThatRPG>
It would be handy if we had a testing library that had the option to run tests in a different lisp process so we could do things that damage a running image...
<NotThatRPG>
(here messing with the package that alexandria uses would be the damage)
<ixelp>
Partial Solution to Testing Challenge - Pastebin.com
morganw has joined #commonlisp
tyson2 has joined #commonlisp
minion has quit [Remote host closed the connection]
specbot has quit [Remote host closed the connection]
cmack has quit [Remote host closed the connection]
czy has quit [Remote host closed the connection]
<NotThatRPG>
Phooey -- uninterning followed by shadowing does NOT do what I expect...
defaultxr has quit [Ping timeout: 276 seconds]
igemnace has quit [Remote host closed the connection]
<NotThatRPG>
Weirdly, seems to alternate b/w working and not working...
simendsjo has joined #commonlisp
defaultxr has joined #commonlisp
attila_lendvai has quit [Read error: Connection reset by peer]
Cymew has joined #commonlisp
gilberth has joined #commonlisp
attila_lendvai has joined #commonlisp
<NotThatRPG>
Do the comparison values for RT have to be literals? Or can they be expressions to evaluate? I believe the answer is no, but wanted to check...
tyson2 has quit [Remote host closed the connection]
waleee has joined #commonlisp
random-nick has quit [Ping timeout: 268 seconds]
random-nick has joined #commonlisp
Gleefre has quit [Remote host closed the connection]
Gleefre has joined #commonlisp
kevingal has joined #commonlisp
Cymew has quit [Ping timeout: 240 seconds]
ym has quit [Ping timeout: 248 seconds]
cage has quit [Quit: rcirc on GNU Emacs 28.2]
triffid has quit [Remote host closed the connection]