2022-07-07 GnuCash IRC logs

00:22:44 *** NoobAlice has quit IRC
01:18:43 *** sbluhm has joined #gnucash
01:25:50 *** storyjesse has quit IRC
01:37:38 *** Gandalf has joined #gnucash
01:59:26 *** fell has quit IRC
02:00:46 *** fell has joined #gnucash
02:00:46 *** ChanServ sets mode: +o fell
02:12:48 *** Aussie_matt has joined #gnucash
02:33:00 *** gjanssens has joined #gnucash
02:33:01 *** ChanServ sets mode: +o gjanssens
02:50:23 *** Hamaryns has joined #gnucash
02:50:23 *** ChanServ sets mode: +v Hamaryns
03:05:20 *** bertbob has joined #gnucash
03:05:20 *** ChanServ sets mode: +v bertbob
04:06:55 *** FH_thecat has joined #gnucash
04:35:23 *** FH_thecat has quit IRC
05:00:17 *** Hamaryns has quit IRC
05:21:34 *** bertbob has quit IRC
05:22:51 *** bertbob has joined #gnucash
05:22:51 *** ChanServ sets mode: +v bertbob
05:23:30 *** Hamaryns has joined #gnucash
05:23:30 *** ChanServ sets mode: +v Hamaryns
05:26:46 *** jmdaemon has quit IRC
06:49:43 *** Hamaryns has quit IRC
07:02:22 *** fell has quit IRC
07:02:52 *** fell has joined #gnucash
07:02:53 *** ChanServ sets mode: +o fell
07:13:02 *** Gandalf1 has joined #gnucash
07:13:30 *** Gandalf has quit IRC
07:13:30 *** Gandalf1 is now known as Gandalf
07:39:39 *** paul has joined #gnucash
08:42:01 *** ChanServ sets mode: +qo warlord warlord
08:42:03 *** warlord sets mode: +o gncbot
10:38:50 *** Aussie_matt has quit IRC
11:21:27 *** sbluhm has quit IRC
11:23:41 *** Gandalf has quit IRC
11:23:56 *** Gandalf has joined #gnucash
11:26:03 *** sbluhm has joined #gnucash
11:34:05 *** TownsendHardware has quit IRC
11:53:04 *** sbluhm has quit IRC
11:58:33 *** ArtGravity has joined #gnucash
11:58:34 *** ChanServ sets mode: +v ArtGravity
12:18:35 *** Gandalf has quit IRC
12:21:35 *** sbluhm has joined #gnucash
12:22:00 *** TownsendHardware has joined #gnucash
12:41:50 *** sbluhm has quit IRC
12:54:41 *** lmat has quit IRC
13:01:28 *** lmat has joined #gnucash
13:30:44 *** guak has joined #gnucash
13:46:29 *** sbluhm has joined #gnucash
13:52:30 *** sbluhm has quit IRC
13:53:43 *** sbluhm has joined #gnucash
14:11:24 <jralls> fell, re gnucash-on-osx docs module: gnucash-git and gnucash-docs-git both default to 'master' but that can be overridden in jhbuildrc-custom. Those modules aren't used for CI or distribution.
14:12:30 <fell> OK
14:12:57 <fell> What about Linas' mail?
14:14:23 <fell> I suspect someone is creating bad bad lang cookies.
14:15:25 <jralls> fell, I don't think our code does anything with cookies.
14:16:51 <jralls> Some of it looks like somebody's probing for an injection vulnerability. The fact that it's logging errors means that our defenses are working, but are creating noise that irritates linas.
14:19:06 <fell> We assume the LANG Cookie was written by us and run no checks on it.
14:23:10 <fell> we could use the one or other PHP Locale:: function
14:29:22 <fell> The lowercase country codes come from 78: strtolower($_SERVER['HTTP_ACCEPT_LANGUAGE'])
14:38:54 *** Gandalf has joined #gnucash
14:48:29 <jralls> fell, LANG is a POST parameter not a cookie. For the rest, I replied to linas's email.
14:55:13 *** sbluhm has quit IRC
15:07:35 *** fabior has joined #gnucash
15:11:23 *** jmdaemon has joined #gnucash
15:22:35 *** ArtGravity has quit IRC
15:23:19 *** fabior has quit IRC
15:23:54 *** ArtGravity has joined #gnucash
15:23:55 *** ChanServ sets mode: +v ArtGravity
15:27:12 *** Cork has quit IRC
15:28:05 <jralls> gjanssens, did you test `ninja install html` with your default on linux? I think warlord *does* need to set -DWITH_HTML_INSTALL=YES in the nightly build scripts, otherwise the contents of https://code.gnucash.org/docs won't get updated.
15:30:39 <warlord> The code currently runs: cmake -DCMAKE_INSTALL_PREFIX=$prefix -DWITH_MOBI=ON $srcdir ; make html pdf epub mobi ; tar -C share/doc -c ... | tar -C <website> -x
15:32:02 *** Cork has joined #gnucash
15:34:55 <jralls> warlord, if it tars from the build dir then it's OK as-is.
15:39:10 <jralls> But I think what gjanssens really wanted was to set EXCLUDE_FROM_ALL to the install command in AddHTMLTarget. That stops ninja install from creating the empty html directories that fell complained about but still allows `ninja install html` to work without setting WITH_HTML_INSTALL.
15:40:47 <warlord> So long as the code I wrote here still works, I'll stay out of the argument :)
15:42:24 *** fabior has joined #gnucash
15:53:48 <gjanssens> jralls: I didn't test 'ninja install html'. I thought I had simply disabled it by default so I figured install html wouldn't work. And on my system it doesn't ?
15:53:48 <gncbot> gjanssens: Sent 18 hours and 17 minutes ago: <chris> good points; this is why instead of trying heuristics every call, it may be safer to scrub txn_type once, set a flag, and show warnings only once.
15:54:13 <gjanssens> (I ran that test now albeit with make rather than ninja)
15:55:16 <jralls> So why disable it? If the user says `ninja install html` wouldn't you think they'd expect it to install the html docs?
16:00:25 *** fabior has quit IRC
16:02:48 <gjanssens> If I remember correctly from when I initially ported the docs build to cmake html was not an install target like it is a build target. So from that I concluded 'ninja install html' didn't work as you'd expect.
16:04:13 <gjanssens> With EXCLUDE_FROM_ALL html can only be installed as part of a components based installation. I have to retest what 'components' actually mean in this context, but it's not the same as a build target for sure.
16:04:34 <gjanssens> That drove me in the direction of making install for html optional via a cmake switch.
16:04:40 <jralls> It is the same as a build target. I've already tested it.
16:05:03 <fell> components={manual|gide}
16:06:04 <jralls> IOW, if EXCLUDE_FROM_ALL is included in the html install parameters, `ninja install` doesn't install html and `ninja install html` does.
16:19:05 <gjanssens> fell: no in cmake components mean a different thing. And the only components currently defined are 'ghelp' and 'chm'. But that's not relevant apparently.
16:19:33 <gjanssens> jralls: if that works I was misguided, so by all means do change as you suggest.
16:20:19 <gjanssens> There is then no need for a cmake switch.
16:20:36 <gjanssens> It will have to go hand in hand with a change in your macos bundle script, as that will now require an explicit html specification for the install rule.
16:20:44 <fell> Perhaps name the cmake components targets?
16:21:34 <gjanssens> cmake calls these components, it would be confusing to name it differently.
16:21:53 <gjanssens> But they aren't relevant apparently so it doesn't matter.
16:22:20 <jralls> gjanssens, No change to gnucash-on-osx required, ninja-args="html" has always been there.
16:22:43 <gjanssens> Ah and that is passed both to build and install ? I didn't know that.
16:23:01 <jralls> Yes, to all ninja commands.
16:23:14 <gjanssens> Cool
16:25:35 *** Gandalf has quit IRC
16:25:49 <gjanssens> That's all for tonight :) Bye
16:25:57 <jralls> That usage of "COMPONENTS" seems unique to `install` and `CMAKE_INSTALL_DEFAULT_COMPONENT_NAME` and isn't explained in either. But if you take "EXCLUDE_FROM_ALL" literally it's easy to interpret as "don't do this for make install all", and testing shows that's what it does.
16:26:04 <jralls> Goodnight, gjanssens.
16:27:06 <gjanssens> Yeah, as I wrote I seem to have misinterpreted all of that. I'm happy there's an easier way.
16:27:16 *** gjanssens has quit IRC
16:29:23 <jralls> fell, back to the locale attacks on the website. I see your comment in lang.php about using Locale::canonicalize, but I don't think that's a good idea. Although the function isn't officially documented (which might by itself make it not a good idea) there's a comment that says it does ICU level 1 and level 2 canonicalization.
16:30:11 <jralls> See https://unicode-org.github.io/icu/userguide/locale/#canonicalization
16:35:07 *** Cork has quit IRC
16:36:23 *** Cork has joined #gnucash
16:37:51 <jralls> I don't think level2 canonicalization makes sense for interpreting URI parameters. Based on the fr-ch snippet in the comment I think you wanted https://www.php.net/manual/en/locale.acceptfromhttp.php anyway.
16:41:24 <fell> If you look at the notes, none of php's locale functions is really good.
16:42:25 <jralls> Yeah. Shall I excise all of that?
16:47:26 *** warlord has quit IRC
16:51:48 <jralls> How about applying the substitution logic in HTTP_ACCEPT_LANGUAGE (which needs to be passed through filter_input) to the lang URL parameter?
17:08:22 <fell> Seems OK to me.
17:25:43 *** ArtGravity has quit IRC
19:06:19 *** ArtGravity has joined #gnucash
19:06:19 *** ChanServ sets mode: +v ArtGravity
19:16:21 *** guak has quit IRC
20:28:07 *** NoobAlice has joined #gnucash
21:27:20 *** Aussie_matt has joined #gnucash
21:47:21 *** mib_u9d9wu has joined #gnucash
21:56:13 <chris> @tell jralls re: internal messaging i think it's hard(TM): too high risk of false positives
21:56:13 <gncbot> chris: The operation succeeded.
21:56:20 <chris> @tell gjanssens re: internal messaging i think it's hard(TM): too high risk of false positives
21:56:20 <gncbot> chris: The operation succeeded.
21:56:20 *** mib_u9d9wu has quit IRC
22:36:21 *** kzootech has joined #gnucash
22:38:12 *** ChanServ sets mode: +v kzootech