Am 20.01.18 16:11 schrieb(en) Albrecht Dreß:
attached is a little patch which improves replying to html-only messages containing national characters. It turns out that html2markdown reliably converts such characters to utf-8 only if an extra command line option is given. As we check for a “zoo” of conversion tools, the configure script verifies that the selected one does actually accept the extra option. Unfortunately, I have no idea how the change should be added to the meson build, so it would be great if someone could have a look at it.
O.k., this seems to be a little complex, as I did not find a way to simply create and erase a file from meson. The recommended (?) way seems to be writing a shell script for that purpose… not really portable! Anyone knows a better way? The meson part is then simple. Attached is patch which shall go on top of the first one, and /should/ detect if the extra command line option is accepted. Unfortunately, I can not build the result, as Ninja is too old on Ubuntu LTS. Cheers, Albrecht.
diff --git a/check_htmlconv.sh b/check_htmlconv.sh
new file mode 100755
index 0000000..a7c26e5
--- /dev/null
+++ b/check_htmlconv.sh
@@ -0,0 +1,7 @@
+#!/bin/sh
+echo "<html/>" > conftest.htm
+$@ conftest.htm
+R=$?
+rm -f conftest.htm
+exit $R
+
diff --git a/meson.build b/meson.build
index c6fc2ae..4cc3a88 100644
--- a/meson.build
+++ b/meson.build
@@ -218,6 +218,11 @@ elif html_widget == 'webkit' or html_widget == 'webkit2'
if html2text.found()
conf.set_quoted('HTML2TEXT', html2text.path(),
description : 'Path to html2text program.')
+ r = run_command('check_htmlconv.sh', html2text.path(), '--unicode-snob')
+ if r.returncode() == 0
+ conf.set('HTML2TEXT_UCOPT', 1,
+ description : 'Defined when the HTML converter accepts the --unicode-snob option.')
+ endif
else
message('You cannot properly quote html-only messags.')
endif
Attachment:
pgpXEq8en2heP.pgp
Description: PGP signature