***DesNoveth_lemur is now known as transformer_lemu
***petercom1and is now known as petercommand
<nalaginrut>I'm testing SQL-injection in the latest Artanis, but seems it's hard to do so, surprise <nalaginrut>I think there're something in guile-dbi, I'm reading it <ArneBab>I’m currently looking into html→epub conversions… <nalaginrut>because guile-dbd-mysql didn't use CLIENT_MULTI_STATEMENTS during init <nalaginrut>so I think dbi can't use multi statements in default <nalaginrut>which means you can't do SQL-injection with multi-statements <nalaginrut>"/test/nala%5c%22;select * from Persons where Lastname=%5c%22hiden_guy" won't work <nalaginrut>well, but I should recommend the author give the flags config in (dbi dbi), let the developers decide how to use <ArneBab>calibre-output looks good - seems like it’ll be easy to turn my py2guile text into an ebook <nalaginrut>unfortunately, kindle is not cool with papers, but good for books <ArneBab>good to know ☺ — I don’t have a kindle, and sadly the tolino with Android sounds like it requires a much faster processor. <nalaginrut>anyway, it can avoid very few kind of SQL-injections... I have to do some work later ***transformer_lemu is now known as monumental_deuce
<taylanub>does `dynamic-link' offer a way to specify a version for the library to be linked? on Debian I just had the problem that the package "libgcrypt11" only provided versioned .so files like .so.11, and (dynamic-link "libgcrypt") failed until I installed libgcrypt11-dev which provides the .so symlink to .so.11. Asking #debian on why the .so symlink isn't in the normal package I got the answer that I <taylanub>shouldn't link to an unversioned .so anyway... <alezost>hello, is it OK to modify %load-path at first and then to use (use-modules ...) in one scm-file? <alezost>I mean I want to load some modules from a calculated path (I can't modify GUILE_LOAD_PATH). How can I do that? <madsy>alezost: I don't think you can easily <alezost>madsy: thanks for the reply. I thought that "(set! %load-path ...) (set! %load-compiled-path ...) (use-modules ...)" would work but I have some strange problems I can't figure out <taylanub>alezost: shouldn't the users be expected to install your modules in whichever standard module directories they want? <alezost>taylanub: no, I want to use modules that override the existing modules from the standard dirs <taylanub>alezost: hrm, OK. it seems one can also just `load' a file that defines a module and it will not change the current module or pollute the current environment, so in place of the %load-path augmentation you could just (load "path/to/my/module.scm") then (use-modules (my module)) <taylanub>I just tested this in the REPL, hope it works in a file too <madsy>alezost: Make sure your paths aren't already built-in existing module paths. I had trouble with that <alezost>taylanub: thank you, your trick with `load' works <alezost>another thing: I've just tried (add-to-load-path "/tmp") and "/tmp" was added to %load-path twice. Is it a bug or it's just me? <taylanub>"‘add-to-load-path’ takes care of modifying the path both at compile-time and at run-time." maybe in the REPL this leads to it happening twice because the compilation and execution happen in the same session *taylanub tests with file <taylanub>on first run, it auto-compiles and thus also adds it twice; on a second run where it uses the precompiled bytecode it adds it once only <taylanub>perhaps it should just generally avoid duplicates? someone want to ask the ML? <taylanub>Elisp's `add-to-list' avoids duplicates. <alezost>Is bug-guile@gnu.org a proper place for reporting about such things? <taylanub>I'd say yes. if it's not considered a bug it will get closed and serve as an archival of this topic being discussed... <taylanub>in fact search first if it's there already :) <sneek> 14:10:05 up 34 days, 14:18, 0 users, load average: 0.00, 0.01, 0.05 <bu^>hello, could gccjit be of any use for guile ? <dsmith-work>I would think eventually, yes! wingo (the main dev working on the compler) mentioned the other day he is not ready to look at it yet. <bu^>ok. seems like it is possible to translate from some bytecode to native, so it could possibly close the chain <taylanub>hrm, there's GCC JIT, and there's GNU Lightning... <bu^>GNU Lightning is maybe less interesting for guile to intermix scheme and C <bu^>libjit uses a VM if I understand properly <bu^>so it's of no interest I think <dsmith-work>Years ago, mvo did someting with GNU Lightning and guile. Back before the move to git. <ijp>sneek: later ask mark_weaver is (equal? (/ 0.0 0.0) (/ 0.0 0.0)) => #t, (= (/ 0.0 0.0) (/ 0.0 0.0)) => #f kosher? <daviid>sorry, i thought i had time for a brief chat but no :) bbl