On 2019-01-13, Scott Kostyshak wrote: > [-- Type: text/plain, Encoding: quoted-printable --]
> On Mon, Jan 07, 2019 at 08:31:03PM -0500, Scott Kostyshak wrote: >> On Wed, Jan 02, 2019 at 06:09:24PM -0500, Scott Kostyshak wrote: >> > For the attached example .lyx file (created from the Embedded Objects >> > manual), consider the following: >> > >> > $ lyx -e lyx16x embedded_mwe.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.16.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.16.16.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.16.16.16.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.16.16.16.16.lyx >> > starting local cmake binary for lyx >> > $ lyx -e lyx16x embedded_mwe.16.16.16.16.16.16.lyx >> > starting local cmake binary for lyx >> > $ sha256sum *lyx >> > 20230fb977921f2c19b11391baffe9fdc929800aecba2ae4ac98585c5d0fb0be >> > embedded_mwe.16.16.16.16.16.16.16.lyx >> > b30cb55c0a67b1e72f441f4d61a69c6152bcb29eadbe4f0c30b8f18ff7b8864e >> > embedded_mwe.16.16.16.16.16.16.lyx >> > 20230fb977921f2c19b11391baffe9fdc929800aecba2ae4ac98585c5d0fb0be >> > embedded_mwe.16.16.16.16.16.lyx >> > b30cb55c0a67b1e72f441f4d61a69c6152bcb29eadbe4f0c30b8f18ff7b8864e >> > embedded_mwe.16.16.16.16.lyx >> > 20230fb977921f2c19b11391baffe9fdc929800aecba2ae4ac98585c5d0fb0be >> > embedded_mwe.16.16.16.lyx >> > b30cb55c0a67b1e72f441f4d61a69c6152bcb29eadbe4f0c30b8f18ff7b8864e >> > embedded_mwe.16.16.lyx >> > 20230fb977921f2c19b11391baffe9fdc929800aecba2ae4ac98585c5d0fb0be >> > embedded_mwe.16.lyx >> > 2c52cc90dcac90c263842fe23f7a8be9b781559714cce245cdd8efc3b212c953 >> > embedded_mwe.lyx >> > >> > Notice how all of the odd-numbered checksums are the same and how all of >> > the even-numbered checksums are the same (except for the non-1.6.x >> > format .lyx file): It seems to alter back and forth between two states. >> > Here is the difference: >> > >> > $ diff -u embedded_mwe.16.lyx embedded_mwe.16.16.lyx >> > --- embedded_mwe.16.lyx 2019-01-02 17:53:39.562138989 -0500 >> > +++ embedded_mwe.16.16.lyx 2019-01-02 17:53:40.254142043 -0500 >> > @@ -51,7 +51,7 @@ >> > <column alignment="center" valignment="top"> >> > <column alignment="center" valignment="top"> >> > <column alignment="center" valignment="top"> >> > -<row endfirsthead="false" caption="true"> >> > +<row caption="true"> >> > <cell multicolumn="1" alignment="center" valignment="top" usebox="none"> >> > \begin_inset Text >> > >> > Is this an issue that should be fixed, or is this expected? >> Attached is a patch. With the patch, the .lyx exports converge. Further, >> the difference in .tex files produced, compared with the original, is >> reduced. >> Without the patch, I get the following: >> $ rm *16* >> $ mylyx master -e lyx16x embedded_mwe.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.16.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.16.16.lyx >> $ md5sum *lyx >> 575ff2f9805e828ecbfaf7cc1d94f154 embedded_mwe.16.16.16.16.lyx >> 86cf64c01d3efa8b4fb73eb790d46e81 embedded_mwe.16.16.16.lyx >> 575ff2f9805e828ecbfaf7cc1d94f154 embedded_mwe.16.16.lyx >> 86cf64c01d3efa8b4fb73eb790d46e81 embedded_mwe.16.lyx >> 1f291ae3ea117832f87aeda6985b347b embedded_mwe.lyx >> $ diff embedded_mwe.16.16.16.lyx embedded_mwe.16.16.16.16.lyx >> 54c54 >> < <row endfirsthead="false" caption="true"> >> --- >> > <row caption="true"> >> $ rm *tex >> $ mylyx master -e pdflatex embedded_mwe.lyx >> $ mylyx master -e pdflatex embedded_mwe.16.16.16.lyx >> $ diff *tex >> 5a6 >> > \synctex=-1 >> 22a24 >> > \endfirsthead >> $ >> Note the difference in .lyx files and of the LaTeX output of \endfirsthead. >> With the patch, I get: >> $ rm *16* >> $ mylyx master -e lyx16x embedded_mwe.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.16.lyx >> $ mylyx master -e lyx16x embedded_mwe.16.16.16.lyx >> $ md5sum *lyx >> 92b43acbf7f54361a9c8a2231037a8ea embedded_mwe.16.16.16.16.lyx >> 92b43acbf7f54361a9c8a2231037a8ea embedded_mwe.16.16.16.lyx >> 92b43acbf7f54361a9c8a2231037a8ea embedded_mwe.16.16.lyx >> 92b43acbf7f54361a9c8a2231037a8ea embedded_mwe.16.lyx >> 1f291ae3ea117832f87aeda6985b347b embedded_mwe.lyx >> $ diff embedded_mwe.16.16.16.lyx embedded_mwe.16.16.16.16.lyx >> $ rm *tex >> $ mylyx master -e pdflatex embedded_mwe.lyx >> $ mylyx master -e pdflatex embedded_mwe.16.16.16.lyx >> $ diff *tex >> 5a6 >> > \synctex=-1 >> $ >> There is still the LaTeX difference of \synctex=-1, but the difference >> of \endfirsthead is gone and there is no difference in .lyx files. >> As a further test, I looked at the difference between the exported .tex >> from 2.1.0 (I don't have 1.6.x) compared to the exported .tex from >> master of the original .lyx file. >> Without the patch: >> $ diff embedded_mwe.16.16.16_2.1.0.tex embedded_mwe.tex >> 1c1 >> < %% LyX 2.1.0 created this file. For more info, see >> http://www.lyx.org/. >> --- >> > %% LyX 2.4.0dev created this file. For more info, see >> https://www.lyx.org/. >> 5a6 >> > \synctex=-1 >> 20c21 >> < \begin{longtable}{|c|c|c|c|c|} >> --- >> > \begin{longtable}[c]{|c|c|c|c|c|} >> 22a24 >> > \endfirsthead >> $ >> With the patch: >> $ diff embedded_mwe.16.16.16_2.1.0.tex embedded_mwe.tex >> 1c1 >> < %% LyX 2.1.0 created this file. For more info, see >> http://www.lyx.org/. >> --- >> > %% LyX 2.4.0dev created this file. For more info, see >> https://www.lyx.org/. >> 5a6 >> > \synctex=-1 >> 20c21 >> < \begin{longtable}{|c|c|c|c|c|} >> --- >> > \begin{longtable}[c]{|c|c|c|c|c|} >> $ >> Note that with the patch, there is no longer the difference of \endfirsthead. >> I arrived at the patch by experimentation and am not sure it's correct. >> But perhaps it gives a clue to someone for what the correct patch would >> be? > I compared .tex output from master-to-1.6.x-to-master roundtrip with and > without the patch to master's direct output, for 11 combinations of > longtable settings. The .lyx files (in current master format) and > corresponding .tex files for the 3 sets (from master and for roundtrip > with 1.6.x with and without the patch) are here: > https://www.dropbox.com/sh/dc4hr8e7652d5w9/AABpDVzb_f-MSYTjFcRkLL7oa > The patch fixes the differences (except for \synctex=-1) for variations > 1, 3, and 4. Variations 2, 10, and 11 have considerable differences for > both with and without the patch, but the differences of 10 and 11 are > strictly smaller with the patch. That is, the patch makes some > improvements without any detected disadvantages. I'm not planning to > look into the differences of 2, 10, and 11, but I can make a trac ticket > if anyone is interested. > I'm planning to commit unless there is an objection. You could do both, commit the partial patch and open a track ticket containing background and open points. Thanks for your work, Günter
