REVAI-4778: Parameterized Debian baseline, static linking with FST

Dockerfile

@@ -1,40 +1,53 @@
+ARG DEBIAN_BASE=bullseye
 # Using kaldi image for pre-built OpenFST, version is 1.7.2
-FROM kaldiasr/kaldi:cpu-debian10-2024-07-29 as kaldi-base
+FROM 414502572119.dkr.ecr.us-west-2.amazonaws.com/kaldi:v_5.5.985_4.0.8_${DEBIAN_BASE} AS kaldi-base
+FROM debian:${DEBIAN_BASE}-slim AS debian-base
+RUN echo "APT::Get::Assume-Yes \"true\";\nAPT::Get::allow \"true\";" | tee -a  /etc/apt/apt.conf.d/90_no_prompt && \
+    echo "APT::Keep-Downloaded-Packages \"false\";" | tee -a  /etc/apt/apt.conf.d/91_no_cache && \
+    apt-get update
 
-FROM debian:11
 
-COPY --from=kaldi-base /opt/kaldi/tools/openfst /opt/openfst
-ENV OPENFST_ROOT /opt/openfst
+FROM debian-base AS stage
+
+COPY --from=kaldi-base /kaldi/tools/openfst /opt/openfst
+ENV OPENFST_ROOT=/opt/openfst
 
 ARG JOBS=4
 
-RUN apt-get update && \
-    apt-get upgrade -y && \
-    apt-get -y install \
+RUN apt-get install \
     cmake \
     g++ \
-    libicu-dev
+    libicu-dev \
+    ccache \
+    git
 
 RUN mkdir /fstalign
 COPY CMakeLists.txt /fstalign/CMakeLists.txt
 COPY src /fstalign/src
 COPY test /fstalign/test
 COPY third-party /fstalign/third-party
 COPY sample_data /fstalign/sample_data
-
+COPY catch2.patch /fstalign/
 WORKDIR /fstalign
-
-RUN mkdir -p /fstalign/build && \
+RUN git apply catch2.patch
+RUN --mount=type=cache,target=/root/.ccache,sharing=locked  \
+    mkdir -p /fstalign/build && \
     cd /fstalign/build && \
     rm -rf * && \
-    cmake .. -DOPENFST_ROOT="${OPENFST_ROOT}" -DDYNAMIC_OPENFST=ON && \
+    cmake .. -DOPENFST_ROOT="${OPENFST_ROOT}" -DDYNAMIC_OPENFST=OFF && \
     make -j${JOBS} VERBOSE=1 && \
     mkdir -p /fstalign/bin && \
     cp /fstalign/build/fstalign /fstalign/bin && \
     strip /fstalign/bin/*
-
+RUN make test
+FROM debian-base AS release
+WORKDIR /fstalign
+RUN if [ "${DEBIAN_BASE}" = "bullseye" ]; then \
+        apt-get install --no-install-recommends libicu67; \
+    else \
+        apt-get install --no-install-recommends libicu72; \
+    fi; \
+    rm -rf /var/lib/apt/lists/*
 COPY tools /fstalign/tools
-
-ENV PATH \
-    /fstalign/bin/:\
-    $PATH
+COPY --from=stage /fstalign/bin   /fstalign/bin
+ENV PATH=/fstalign/bin:$PATH

catch2.patch

@@ -0,0 +1,28 @@
+diff --git a/third-party/catch2/include/internal/catch_fatal_condition.cpp b/third-party/catch2/include/internal/catch_fatal_condition.cpp
+index 24e16d3c..78ddc6e5 100644
+--- a/third-party/catch2/include/internal/catch_fatal_condition.cpp
++++ b/third-party/catch2/include/internal/catch_fatal_condition.cpp
+@@ -97,7 +97,8 @@ namespace Catch {
+
+     // 32kb for the alternate stack seems to be sufficient. However, this value
+     // is experimentally determined, so that's not guaranteed.
+-    constexpr static std::size_t sigStackSize = 32768 >= MINSIGSTKSZ ? 32768 : MINSIGSTKSZ;
++    // Note: MINSIGSTKSZ is not constexpr in glibc 2.34+, so we use a fixed safe value
++    constexpr static std::size_t sigStackSize = 32768;
+
+     static SignalDefs signalDefs[] = {
+         { SIGINT,  "SIGINT - Terminal interrupt signal" },
+diff --git a/third-party/catch2/single_include/catch2/catch.hpp b/third-party/catch2/single_include/catch2/catch.hpp
+index b324e56a..6a104637 100644
+--- a/third-party/catch2/single_include/catch2/catch.hpp
++++ b/third-party/catch2/single_include/catch2/catch.hpp
+@@ -7969,7 +7969,8 @@ namespace Catch {
+
+     // 32kb for the alternate stack seems to be sufficient. However, this value
+     // is experimentally determined, so that's not guaranteed.
+-    constexpr static std::size_t sigStackSize = 32768 >= MINSIGSTKSZ ? 32768 : MINSIGSTKSZ;
++    // Note: MINSIGSTKSZ is not constexpr in glibc 2.34+, so we use a fixed safe value
++    constexpr static std::size_t sigStackSize = 32768;
+
+     static SignalDefs signalDefs[] = {
+         { SIGINT,  "SIGINT - Terminal interrupt signal" },

src/FstFileLoader.h

@@ -3,7 +3,7 @@
  * 
  * FstLoader for loading a serialized FST from disk.
  * 
- * Quinn McNamara (quinn@rev.com)
+ * JP Robichaud (jp@rev.com)
  * 2020
  */
 

src/wer.h

@@ -3,7 +3,7 @@
  *
  * Collection of functions specific to the WER subcommand.
  *
- * Quinn McNamara (quinn@rev.com)
+ * JP Robichaud (jp@rev.com)
  * 2021
  */
 #include "fstalign.h"