Guess input encoding and load case change code earlier (#1090)

... also addresses https://tex.stackexchange.com/q/581409/ Ping back to #1058. Reverts bits of 3165987.
plk · Feb 1, 2021 · 15a62fb · 15a62fb
1 parent 6a089ee
commit 15a62fb
Show file tree

Hide file tree

Showing 2 changed files with 68 additions and 31 deletions.
diff --git a/doc/latex/biblatex/biblatex.tex b/doc/latex/biblatex/biblatex.tex
@@ -1919,7 +1919,7 @@ \subsubsection{Sorting and Encoding Issues}
 
 \paragraph{Specifying Encodings}
 \label{bib:cav:enc:enc}
-When using a non-\acr{US-ASCII} encoding in the \file{bib} file, it is important to understand what \biblatex can do for you and what may require manual intervention. The package takes care of the \latex side, \ie it ensures that the data imported from the \file{bbl} file is interpreted correctly, provided that the \opt{bibencoding} package option (or the datasource specific override for this, see \secref{use:bib:res}) is set properly. All of this is handled automatically and no further steps, apart from setting the \opt{bibencoding} option in certain cases, are required. Here are a few typical usage scenarios along with the relevant lines from the document preamble:
+When using a non-\acr{US-ASCII} encoding in the \file{bib} file, it is important to understand what \biblatex can do for you and what may require manual intervention. The package takes care of the \latex side, \ie it ensures that the data imported from the \file{bbl} file is interpreted correctly, provided that the \opt{bibencoding} package option (or the datasource specific override for this, see \secref{use:bib:res}) is set properly. All of this is handled automatically and no further steps, apart from setting the \opt{bibencoding} option in certain cases (namely when the encoding of the \file{bib} file differs from the encoding of the \file{tex} file), are required provided that you set up your document encoding (\ie load \sty{inputenc} or related packages if required) \emph{before} \biblatex is loaded. Here are a few typical usage scenarios along with the relevant lines from the document preamble:
 
 \begin{itemize}
 \setlength{\itemsep}{0pt}

diff --git a/tex/latex/biblatex/biblatex.sty b/tex/latex/biblatex/biblatex.sty
@@ -206,6 +206,14 @@
   {\global\cslet{blx@pkgloaded@polyglossia}\@empty}
   {}
 
+\@ifpackageloaded{inputenc}
+  {\global\cslet{blx@pkgloaded@inputenc}\@empty}
+  {}
+
+\@ifpackageloaded{inputenx}
+  {\global\cslet{blx@pkgloaded@inputenx}\@empty}
+  {}
+
 % this part is deferred to \AtEndPreamble
 \def\blx@packageincompatibility@endpreambleonly{%
   % people should not be abusing noerroretextools,
@@ -219,6 +227,17 @@
           Please do not define '\string\blx@noerroretextools'\MessageBreak
           unless you really need it}}}
     {}%
+  \@ifpackageloaded{inputenc}
+    {\ifcsundef{blx@pkgloaded@inputenc}
+       {\blx@warning@noline{Load 'inputenc' before biblatex}}
+       {}}
+    {}%
+  \@ifpackageloaded{inputenx}
+    {\ifcsundef{blx@pkgloaded@inputenx}
+       {}
+       {\blx@warning@noline{Load 'inputenx' before biblatex}}
+       {}}
+    {}%
   % polyglossia pretends to be babel, so the nested structure helps
   % to make sure that babel really is babel
   \@ifpackageloaded{polyglossia}
@@ -14409,8 +14428,8 @@
 
 \DeclareBibliographyOption[string]{texencoding}{%
   \ifstrequal{#1}{auto}
-    {\undef\blx@texencoding}
-    {\def\blx@texencoding{#1}}}
+    {\undef\blx@texencoding@fromoption}
+    {\def\blx@texencoding@fromoption{#1}}}
 
 \DeclareBibliographyOption[string]{bibencoding}{%
   \ifcsdef{blx@bibenc@#1}
@@ -15617,7 +15636,6 @@
 % Process load-time options
 \ProcessOptions*
 
-
 % Switch to BibTeX support if requested
 \iftoggle{blx@bibtex}
   {\blx@inputonce{blx-bibtex.def}{BibTeX backend compatibility}{}{}{}{}}
@@ -15640,6 +15658,41 @@
 % Process preamble options
 \blx@processoptions
 
+% first stab at guessing the file encoding
+\def\blx@guesstexencoding{%
+  \ifdef\blx@texencoding@fromoption
+    {\let\blx@texencoding\blx@texencoding@fromoption
+     \blx@info@noline{%
+       Input encoding '\blx@texencoding' specified}}
+    {\ifundef\inputencodingname
+       {\ifundef\XeTeXrevision
+          {\ifundef\luatexversion
+             {\def\blx@texencoding{ascii}%
+              \blx@info@noline{%
+                No input encoding detected.\MessageBreak
+                Assuming '\blx@texencoding'}}
+             {\def\blx@texencoding{utf8}%
+              \blx@info@noline{%
+                LuaTeX detected.\MessageBreak
+                Assuming input encoding '\blx@texencoding'}}}
+          {\def\blx@texencoding{utf8}%
+           \blx@info@noline{%
+             XeTeX detected.\MessageBreak
+             Assuming input encoding '\blx@texencoding'}}}
+       {\ifcsdef{blx@inpenc@\inputencodingname}
+          {\letcs\blx@texencoding{blx@inpenc@\inputencodingname}}
+          {\let\blx@texencoding\inputencodingname}%
+        \blx@info@noline{%
+          Input encoding '\blx@texencoding' detected}}}}
+
+\blx@guesstexencoding
+\let\blx@texencoding@firstguess\blx@texencoding
+
+% this ultimately loads either
+% blx-case-expl3.def or blx-case-latex2e.def
+\blx@deferoption@casechanger
+\blx@casechange@postload
+
 % Deferred setup at end of preamble
 \AtEndPreamble{%
   \blx@checkoptions@global
@@ -15720,30 +15773,18 @@
     \abx@dodates
     \do{labeldate}}
 
+
+
 \def\blx@checkencoding{%
-  \ifdef\blx@texencoding
-    {\blx@info@noline{%
-       Input encoding '\blx@texencoding' specified}}
-    {\ifundef\inputencodingname
-       {\ifundef\XeTeXrevision
-          {\ifundef\luatexversion
-             {\def\blx@texencoding{ascii}%
-              \blx@info@noline{%
-                No input encoding detected.\MessageBreak
-                Assuming '\blx@texencoding'}}
-             {\def\blx@texencoding{utf8}%
-              \blx@info@noline{%
-                LuaTeX detected.\MessageBreak
-                Assuming input encoding '\blx@texencoding'}}}
-          {\def\blx@texencoding{utf8}%
-           \blx@info@noline{%
-             XeTeX detected.\MessageBreak
-             Assuming input encoding '\blx@texencoding'}}}
-       {\ifcsdef{blx@inpenc@\inputencodingname}
-          {\letcs\blx@texencoding{blx@inpenc@\inputencodingname}}
-          {\let\blx@texencoding\inputencodingname}%
-        \blx@info@noline{%
-          Input encoding '\blx@texencoding' detected}}}%
+  \blx@guesstexencoding
+  \ifdefstrequal\blx@texencoding@firstguess\blx@texencoding
+    {}
+    {\blx@warning@noline{%
+      Initial guess for file encoding (\blx@texencoding@firstguess)\MessageBreak
+      and new guess (\blx@texencoding) do not match.\MessageBreak
+      Please set up your document encoding *before*\MessageBreak
+      loading biblatex.\MessageBreak
+      The case changing code might not work as expected}}%
   \ifdef\blx@bibencoding
     {\ifdefstring\blx@bibencoding{ascii}
        {\blx@info@noline{%
@@ -15782,10 +15823,6 @@
 \AtBeginDocument{%
   \blx@auxwrite\@mainaux{}{\string\abx@aux@refcontext{\blx@refcontext@context}}
   \blx@checkencoding
-  % this ultimately loads either
-  % blx-case-expl3.def or blx-case-latex2e.def
-  \blx@deferoption@casechanger
-  \blx@casechange@postload
   \if@filesw
     \blx@ctrlwrite
   \fi