On 9/13/22 18:48, Udicoudco wrote:
On Tue, Aug 30, 2022 at 10:27 AM Jürgen Spitzmüller<jspi...@gmail.com>  wrote:
Am Donnerstag, dem 14.07.2022 um 14:17 +0300 schrieb Udicoudco:
Hello again,
Hi,
Hi Jürgen,
Thank you for the reply.

Thanks for this and sorry for the late reply. Most people are busy with
their real lifes.
No apologies needed, thank  you all for developing great software.

I had some mistakes in the .inc file and the .lyx file. The updated
files are attached.
As I do not use theorems myself, I cannot comment on the usage and
usefulness of thmtools and the need for a module. So you probably have
to be even more patient until the math users are back.
Well, the matter is not urgent, and I have patience :).

One remark on the layout:

AddToPreamble
          \usepackage{amsthm}
          \usepackage{thmtools}
EndPreamble

At least for amsthm, this should rather be

Require amsthm
I've fixed this issue (see the attached file). I'v also used the
DependsOn key so that the module will call thmtools only once.

I think also thmtools should be added to the packages LyX knows and
loaded that way.
I don't know how to do that.

Regards,
Udi

Hi Udi,

I've been playing a bit with your module, and I've encountered a couple of non-fatal issues. The first may sound a bit silly. If you have a list of theorems but do not have any theorem-like environments (either because you put in the list of theorems code first or because you deleted all the things that would have been listed and have not yet added new ones), the document will not compile because the thm-tools package has not been loaded. Either requiring thm-tools in the list of theorems environment or making it depend on theorems (the way claim, lemma etc. do) should cure that.

The second one is perhaps a bug in the thm-tools package itself. To reproduce it, you can follow these steps.

1. Open your test_thmtools_module.lyx file and compile it (which works,
   of course).
2. Delete the claims in section 2 (so that section 2 begins with
   Conjecture 2.1) and recompile. When I do this, I get four undefined
   control sequence errors, at least three of which contain the macro
   \claimname. If I select "view output anyway", the document seems to
   display correctly.
3. Now it gets goofier. If I make some trivial text edit (just to force
   a recompilation), say adding a character to the section heading, and
   recompile, the document compiles correctly without the error messages.

Something similar happens if, in step 2, I delete all the definitions rather than all the claims, except that the errors now refer to \definitionname. It's important to note that this only happens if I compile the original document first. If I open the document, delete claims or definitions, and then compile for the first time, there is no error. So something is carrying over (in the buffer directory) from the initial compilation that maybe ought not carry over. Also, it only seems to happen when I delete all of one type of environment (all claims, all definitions, ...). As long as I leave one claim or one definition intact, the error does not occur.

I did a little digging. It turns out that compilation creates both a .aux file and a .loe file in the buffer directory. After compiling the full document, both contain instances of \claimname. After deleting all claims and recompiling (producing the error messages), both are rebuilt with no references to \claimname. If, however, I delete either one after compiling the full document and before compiling the edited version, the edited version generates no error messages. So I suspect there is some sort of timing issue here, in which the first compilation of the edited document inherits some undefined instances of \claimname (missing from both .aux and .loe after they were rebuilt) whereas a recompilation does not inherit \claimname. Deleting either the .aux or .loe file may force them to rebuilt before \claimname leaks into the main document. I hope that makes sense. (If it does, please explain to me what I just wrote. :-))

Cheers,

Paul
-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel

Reply via email to